Cap 13 Radar
EDWARD VELASCO
2024-02-07
pck <- (c("tidyr", "dplyr", "readxl", "ggplot2", "randomForest", "car", "Metrics","raster","rasterVis","rgdal", "mapview","RColorBrewer","ggplot2", "sf","glcm", "ggpubr", "sf", "corrplot", "caret"))
library(viridis)
library(ggpointdensity)
library(ggplot2)
library(ggmap)
library(dplyr)
library(spatstat)
library(spatialEco)
library(usdm)
sapply(pck, require, character.only=TRUE)## tidyr dplyr readxl ggplot2 randomForest car
## TRUE TRUE TRUE TRUE TRUE TRUE
## Metrics raster rasterVis rgdal mapview RColorBrewer
## TRUE TRUE TRUE TRUE TRUE TRUE
## ggplot2 sf glcm ggpubr sf corrplot
## TRUE TRUE TRUE TRUE TRUE TRUE
## caret
## TRUEPrimero cargaremos las variables SAR ALOS 2, las cuales se descargaron previamente con el código de Google earth engine (https://code.earthengine.google.com/?scriptPath=users%2Feavelascop1%2Feavelascop%3AALOS_2%20Download).
El mosaico global PALSAR/PALSAR-2 de 25 m es una imagen SAR global perfecta creada mediante el mosaico de franjas de imágenes SAR de PALSAR/PALSAR-2. para el dataset de GEE Las imágenes SAR se ortorectificaron y se corrigió la pendiente utilizando el modelo de superficie digital ALOS World 3D - 30 m (AW3D30).
adicional mente en el codigo proporcionado para la descarga del mosaico anual de ALOS 2 se aplico un filtro de de “SPECKLE” para reducir el moteado característico de las imágenes SAR
getwd()## [1] "D:/cap_13_Radar"alos2 <- stack(list.files(path="D:/cap_13_Radar/ALOS_15_SPK",full.names=TRUE))
plot(alos2)
names(alos2)## [1] "HH" "HV"plot(alos2)
como los valores de las Imagenes ALOS 2 vienen por efecto con valores de dNúmeros digitales (DN) de 16 bits. Los valores de DN se pueden convertir a valores gamma cero en unidades de decibelios (dB) usando la siguiente ecuación:
γ₀ = 10log₁₀(DN²) - 83.0 dB
alos2_Gamma_dB <- 10 * log10 (alos2^2) - 83.0
plot(alos2_Gamma_dB )
A continuación, convierta los valores de dB a retrodispersión de
potencia aplicando la siguiente ecuación:
Gamma_pw = 10^(0,1*Gamma_dB)
alos2_Gamma_pw <- 10^(0.1*alos2_Gamma_dB)
names(alos2_Gamma_pw)## [1] "HH" "HV"plot(alos2_Gamma_pw)
ahora procederemos a cargar el dataset SAR sentinel 1 que fueron
descargados utilizando el script para google earth engine (https://code.earthengine.google.com/?scriptPath=users%2Feavelascop1%2Feavelascop%3ASENTINEL1)
S1_2015 <- stack(list.files(path="D:/cap_13_Radar/S1_2015_SPK",full.names=TRUE))
names(S1_2015) <- c("vhAscDesc_2015", "vhAsc_2015","vhDesc_2015" ,"vvAscDesc_2015", "vvAsc_2015","vvDesc_2015")
#resampleamos al tamaño de pixel de las imagenes ALOS2
S1_2015 <- resample(S1_2015, alos2_Gamma_pw)
plot(S1_2015)
names(S1_2015)## [1] "vhAscDesc_2015" "vhAsc_2015" "vhDesc_2015" "vvAscDesc_2015"
## [5] "vvAsc_2015" "vvDesc_2015"los datos del dataset de sentinel 1 de GEE vienen en valores de DB, por lo que para comvertirlos en valores de Gamma PW debemos utilizar la formula usada anteriormente con los datos de Alos ”
S1_2015_Gamma_pw <- 10^(0.1*S1_2015)
plot(S1_2015_Gamma_pw)
names_S1_2015 <- names(S1_2015_Gamma_pw)carga indices Landsat
Los datos SAR funcionan mejor cuando se combinan con otro tipo de sensores ópticos como lidar o imágenes ópticas satelitales, en nuestro caso de estudio utilizaremos datos opticos ya que las colecciones son libres y tienen cobertura mundial, por lo que a continuación haremos la carga de los indices satelitales landsat 8 calculados previamente en el codigo de google earth engine: https://code.earthengine.google.com/?noload=1&scriptPath=users%2Feavelascop1%2FCAP_13_RADAR%3AINDEX_L8
LANDSAT <- stack(list.files(path="D:/cap_13_Radar/LANDSAT_2015",full.names=TRUE))
LANDSAT <- resample(LANDSAT, alos2_Gamma_pw)
plot(LANDSAT)
Cargamos los datos de las parcelas de campo, las cuales se encuentra en un shp denominado inventario_biomasawgs84.shp
library(sf)
getwd()## [1] "D:/cap_13_Radar"parcelas <- st_read("D:/cap_13_Radar/inventario_biomasa_wgs84/inventario_biomasa_wgs84.shp")## Reading layer `inventario_biomasa_wgs84' from data source
## `D:\cap_13_Radar\inventario_biomasa_wgs84\inventario_biomasa_wgs84.shp'
## using driver `ESRI Shapefile'## Warning in CPL_read_ogr(dsn, layer, query, as.character(options), quiet, : GDAL
## Error 1: PROJ: proj_identify: C:\Program
## Files\PostgreSQL\14\share\contrib\postgis-3.1\proj\proj.db lacks
## DATABASE.LAYOUT.VERSION.MAJOR / DATABASE.LAYOUT.VERSION.MINOR metadata. It
## comes from another PROJ installation.## Simple feature collection with 2107 features and 11 fields
## Geometry type: POINT
## Dimension: XY
## Bounding box: xmin: -2.865975 ymin: 37.14839 xmax: -2.193143 ymax: 37.35033
## Geodetic CRS: WGS 84plot(parcelas[3])
names(parcelas)## [1] "FID" "OBJECTI" "N_Cintf" "Dnm_cm" "Npies_h" "AB_m2_h"
## [7] "Dmc_cm" "Ho_m" "Bttl_Kg" "MgC.pr_" "Ctt_MC." "geometry"Revisamos y limpiamos los datos de las parcelas
# Estudio previo de los datos
# Se carga la librería 'car' para usar algunas funciones útiles.
library(car)
# Se imprime un resumen estadístico de la variable 'Ctt_MC.' en el dataframe 'parcelas'.
summary(parcelas$Ctt_MC.)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.00 11.18 29.09 37.76 57.55 331.41# Se crea un diagrama de caja (boxplot) para visualizar la distribución de la variable 'Ctt_MC.' en 'parcelas'.
Boxplot(parcelas$Ctt_MC.)
## [1] 779 838 847 996 1752 1073 1278 1617 1031 1213# Se identifican los valores atípicos (outliers) de la variable 'Ctt_MC.' utilizando boxplot.
outliners <- boxplot(parcelas$Ctt_MC.)$out
outliners## [1] 331.4067 250.3730 196.3495 130.4630 127.2519 127.9674 131.3214 128.0279
## [9] 149.0483 132.7146 138.2332 132.4152 135.2561 128.2277 129.4878 134.6589
## [17] 141.0414# Se filtran los datos para mantener solo aquellos cuyo valor en 'Ctt_MC.' sea menor o igual a 127.2519.
parcelas <- subset.data.frame(parcelas, parcelas$Ctt_MC. <= 120 )
# Se imprime un resumen estadístico de 'Ctt_MC.' después de filtrar los valores.
summary(parcelas$Ctt_MC.)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.00 11.05 28.55 36.42 55.87 119.38# Se crea un nuevo boxplot para visualizar la distribución actualizada de 'Ctt_MC.'.
Boxplot(parcelas$Ctt_MC.)
# Se imprime un resumen estadístico adicional de 'Ctt_MC.' después del filtrado.
summary(parcelas$Ctt_MC.)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.00 11.05 28.55 36.42 55.87 119.38# Se crea un histograma para visualizar la distribución de 'Ctt_MC.' después del filtrado.
hist(parcelas$Ctt_MC.)
# Se identifican los índices de las observaciones cuyo valor en 'Ctt_MC.' es igual a cero.
which(parcelas$Ctt_MC. == 0)## [1] 696 786 792 820 828# Se eliminan las observaciones donde 'Ctt_MC.' es igual a cero del dataframe 'parcelas'.
parcelas <- parcelas[-which(parcelas$Ctt_MC. == 0),]
# Se crea otro histograma para visualizar la distribución de 'Ctt_MC.' después de eliminar los valores iguales a cero.
hist(parcelas$Ctt_MC.)
# Se imprime un resumen estadístico de la variable 'Ctt_MC.' en el dataframe 'parcelas'.
summary(parcelas$Bttl_Kg)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 4.819 235.562 449.461 564.605 817.806 5540.780# Se crea un diagrama de caja (boxplot) para visualizar la distribución de la variable 'Ctt_MC.' en 'parcelas'.
Boxplot(parcelas$Bttl_Kg)
## [1] 61 1296 607 1290 1524 1505 612 1929 601 1930names(parcelas)## [1] "FID" "OBJECTI" "N_Cintf" "Dnm_cm" "Npies_h" "AB_m2_h"
## [7] "Dmc_cm" "Ho_m" "Bttl_Kg" "MgC.pr_" "Ctt_MC." "geometry"Ahora vamos a cargar el DEM derivado del la colección ALOS World 3D - 30m (AW3D30), la cual es un conjunto de datos de modelo de superficie digital (DSM) global con una resolución horizontal de aproximadamente 30 metros (malla de 1 arco segundo).
El script de descarga de GEE se encuentra en: https://code.earthengine.google.com/?scriptPath=users%2Feavelascop1%2Feavelascop%3Adem
getwd()## [1] "D:/cap_13_Radar"# Cargar el DEM
DEM <- raster("D:/cap_13_Radar/DEM/DEM_ALOS.tif")
# Mostrar el mapa de aspecto
plot(DEM , main = "MAPA DE ELEVACIONES DEM")
# Calcular el aspect
ASPECT <- terrain(DEM, opt = "aspect")
# Mostrar el mapa de aspecto
plot(ASPECT, main = "Mapa de Aspecto")
# Calcular la pendiente
SLOPE <- terrain(DEM, opt = "slope")
# Mostrar el mapa de pendiente
plot(SLOPE, main = "Mapa de Pendiente")
TOPO <- stack(DEM, SLOPE, ASPECT)
TOPO <- resample(TOPO, alos2_Gamma_pw)
plot(TOPO)
names(TOPO) <- c("DEM", "SLOPE", "ASPECT")Cargamos los polígonos de la delimitacion de los pinares en filabres por especie
pinus <- rgdal::readOGR("D:/cap_13_Radar/inventario_biomasa_wgs84/pinos_filabres_wgs_84.shp")## Warning: OGR support is provided by the sf and terra packages among others## Warning: OGR support is provided by the sf and terra packages among others## Warning: OGR support is provided by the sf and terra packages among others## Warning: OGR support is provided by the sf and terra packages among others## Warning: OGR support is provided by the sf and terra packages among others## Warning: OGR support is provided by the sf and terra packages among others## Warning: OGR support is provided by the sf and terra packages among others## OGR data source with driver: ESRI Shapefile
## Source: "D:\cap_13_Radar\inventario_biomasa_wgs84\pinos_filabres_wgs_84.shp", layer: "pinos_filabres_wgs_84"
## with 1661 features
## It has 4 fieldscrs(pinus) <- "+proj=longlat +datum=WGS84"
mapview(pinus) # Cargar la librería sf
library(sf)
# Filtrar por especie "PH"
pinus_PH <- pinus[pinus$ESPECIE == "PH", ]
pinus_PP <- pinus[pinus$ESPECIE == "PT", ]extraemos los puntos de coincidencia entre las parcelas y los rasters preparados anteriormente
#stack final de los set ya preparados
rasters1 <- stack(alos2_Gamma_pw, S1_2015_Gamma_pw,LANDSAT, TOPO)
plot(rasters1)
#hacemos el extract con un buffer de 50 metros para minimizar los posibles errores de reproyeccion de coordenadas
rasters1_extract <- raster::extract(rasters1,parcelas, buffer=50, fun=mean)
#extraemos los puntos
MALLA_SAR_50<-cbind(parcelas,rasters1_extract)
write.csv(MALLA_SAR_50, "MALLA_SAR_50.csv")
plot(MALLA_SAR_50, max.plot = 50)
names(MALLA_SAR_50)## [1] "FID" "OBJECTI" "N_Cintf" "Dnm_cm"
## [5] "Npies_h" "AB_m2_h" "Dmc_cm" "Ho_m"
## [9] "Bttl_Kg" "MgC.pr_" "Ctt_MC." "HH"
## [13] "HV" "vhAscDesc_2015" "vhAsc_2015" "vhDesc_2015"
## [17] "vvAscDesc_2015" "vvAsc_2015" "vvDesc_2015" "SR_B2"
## [21] "brightness" "DVI" "EVI" "GNDVI"
## [25] "SR_B3" "greenness" "GVI" "LAI"
## [29] "NBR" "NDBI" "NDVI" "SR_B5"
## [33] "NLI" "NRVI" "OSAVI" "RDVI"
## [37] "SR_B4" "RVI" "SAVI" "SR_B6"
## [41] "SR_B7" "SLAVI" "TVI" "wetness"
## [45] "DEM" "SLOPE" "ASPECT" "geometry"ds <- as.data.frame(MALLA_SAR_50)
ds <- ds[-48]
#filtramos el dataset a nuestras especies de interés
DS_PH <- filter(ds, N_Cintf == "Pinus halepensis")
DS_PP <- filter(ds, N_Cintf == "Pinus pinaster") ### Calcular VIF basandonos en las funciones vifcor y vifstep
#############################################################
var.df <- as.data.frame(rasters1)
v.1 <-vifcor(var.df , th=0.8) # Busca un par de variables que tengan la maxima correlacion lineal (mayor que th) y excluye una de ellas que tenga mayor VIF. El procedimiento se repite hasta que no quede ninguna variable con alto coeficiente de correlacion (mayor que el umbral) con otras variables## Warning in summary.lm(lm(as.formula(paste(colnames(y)[w[i]], "~.", sep = "")),
## : essentially perfect fit: summary may be unreliable
## Warning in summary.lm(lm(as.formula(paste(colnames(y)[w[i]], "~.", sep = "")),
## : essentially perfect fit: summary may be unreliablev.1## 24 variables from the 36 input variables have collinearity problem:
##
## NDBI LAI SAVI NDVI SLAVI OSAVI TVI GNDVI NLI RVI SR_B6 RDVI SR_B4 SR_B2 brightness DVI greenness vhAsc_2015 GVI vhDesc_2015 SR_B7 SR_B5 vvAscDesc_2015 NBR
##
## After excluding the collinear variables, the linear correlation coefficients ranges between:
## min correlation ( NRVI ~ vvAsc_2015 ): -0.002000741
## max correlation ( NRVI ~ HV ): -0.6911023
##
## ---------- VIFs of the remained variables --------
## Variables VIF
## 1 HH 1.812146
## 2 HV 3.786369
## 3 vhAscDesc_2015 2.534848
## 4 vvAsc_2015 1.995375
## 5 vvDesc_2015 1.943983
## 6 EVI 1.040558
## 7 SR_B3 1.853545
## 8 NRVI 2.530722
## 9 wetness 1.487489
## 10 DEM 1.564542
## 11 SLOPE 1.669587
## 12 ASPECT 1.264817v.2 <- vifstep(var.df, th=7) # Calcula el VIF para todas las variables, excluye una con el VIF mas alto (mayor que el umbral), repite el procedimiento hasta que no quede ninguna variable con VIF mayor que th.## Warning in summary.lm(lm(.dd[, i] ~ ., data = .dd[-i])): essentially perfect
## fit: summary may be unreliable## Warning in summary.lm(lm(.dd[, i] ~ ., data = .dd[-i])): essentially perfect
## fit: summary may be unreliablev.2## 22 variables from the 36 input variables have collinearity problem:
##
## NBR EVI SAVI NDVI greenness SR_B4 SR_B6 OSAVI TVI SLAVI brightness RDVI NLI GNDVI SR_B5 SR_B3 RVI NRVI wetness vhAscDesc_2015 GVI SR_B7
##
## After excluding the collinear variables, the linear correlation coefficients ranges between:
## min correlation ( LAI ~ vvDesc_2015 ): -0.003250217
## max correlation ( vvAsc_2015 ~ vhAsc_2015 ): 0.8879027
##
## ---------- VIFs of the remained variables --------
## Variables VIF
## 1 HH 2.029763
## 2 HV 3.561361
## 3 vhAsc_2015 6.254103
## 4 vhDesc_2015 4.725050
## 5 vvAscDesc_2015 2.841023
## 6 vvAsc_2015 6.122730
## 7 vvDesc_2015 4.029359
## 8 SR_B2 1.425038
## 9 DVI 2.460300
## 10 LAI 1.010155
## 11 NDBI 2.685220
## 12 DEM 1.529738
## 13 SLOPE 1.628671
## 14 ASPECT 1.330869re1 <- exclude(rasters1,v.2)
re1## class : RasterStack
## dimensions : 1794, 3804, 6824376, 14 (nrow, ncol, ncell, nlayers)
## resolution : 0.0002245788, 0.0002245788 (x, y)
## extent : -3.009356, -2.155058, 37.09211, 37.49501 (xmin, xmax, ymin, ymax)
## crs : +proj=longlat +datum=WGS84 +no_defs
## names : HH, HV, vhAsc_2015, vhDesc_2015, vvAscDesc_2015, vvAsc_2015, vvDesc_2015, SR_B2, DVI, LAI, NDBI, DEM, SLOPE, ASPECT
## min values : 2.432126e-03, 2.669044e-04, 3.259987e-03, 3.138417e-03, 1.276776e-02, 8.352074e-03, 7.183256e-03, 4.916208e+03, -1.608826e+03, -1.607692e+04, -4.465110e-01, 3.307917e+02, 0.000000e+00, 0.000000e+00
## max values : 1.159834e+01, 1.705286e+00, 9.583676e-01, 1.342883e+00, 1.262230e+01, 1.772115e+01, 9.253421e+01, 2.698925e+04, 1.862317e+04, 1.346486e+04, 1.724137e-01, 2.605000e+03, 1.141819e+00, 6.283185e+00names(re1)## [1] "HH" "HV" "vhAsc_2015" "vhDesc_2015"
## [5] "vvAscDesc_2015" "vvAsc_2015" "vvDesc_2015" "SR_B2"
## [9] "DVI" "LAI" "NDBI" "DEM"
## [13] "SLOPE" "ASPECT"plot(re1)
Ahora dividiremos el set de datos en 80% de datos para entrenamiento y
20% para testeo para los dos datasets de Pinus pinaster y Pinus
halepensis
library(caret)
# Dividir los datos en un conjunto de entrenamiento (80%) y un conjunto de prueba 20%) para PH
set.seed(123)
train_index <- createDataPartition(y =DS_PH$Ctt_MC., p = 0.8, list = FALSE)
train_PH <-DS_PH[train_index, ]
test_PH <-DS_PH[-train_index, ]
summary(train_PH)## FID OBJECTI N_Cintf Dnm_cm
## Min. : 21.0 Min. : 3010 Length:506 Min. :10.61
## 1st Qu.: 943.2 1st Qu.: 3950 Class :character 1st Qu.:13.72
## Median :1668.5 Median : 4676 Mode :character Median :15.32
## Mean :1465.2 Mean : 8160 Mean :17.30
## 3rd Qu.:1864.8 3rd Qu.: 4872 3rd Qu.:19.17
## Max. :2839.0 Max. :46696 Max. :58.37
## Npies_h AB_m2_h Dmc_cm Ho_m
## Min. : 5.093 Min. : 0.1772 Min. :10.61 Min. : 3.546
## 1st Qu.: 220.547 1st Qu.: 4.0819 1st Qu.:13.88 1st Qu.: 6.851
## Median : 380.945 Median : 7.1103 Median :15.63 Median : 7.922
## Mean : 416.825 Mean :10.1075 Mean :17.60 Mean : 9.116
## 3rd Qu.: 541.343 3rd Qu.:14.4820 3rd Qu.:19.69 3rd Qu.:11.005
## Max. :2260.002 Max. :42.5279 Max. :60.07 Max. :27.484
## Bttl_Kg MgC.pr_ Ctt_MC. HH
## Min. : 129.3 Min. :0.01264 Min. : 0.2534 Min. :0.0159
## 1st Qu.: 224.1 1st Qu.:0.33477 1st Qu.: 6.7123 1st Qu.:0.1125
## Median : 338.5 Median :0.60418 Median :12.1139 Median :0.1505
## Mean : 528.2 Mean :0.98720 Mean :19.7935 Mean :0.1703
## 3rd Qu.: 661.4 3rd Qu.:1.37863 3rd Qu.:27.6418 3rd Qu.:0.2040
## Max. :5540.8 Max. :4.89050 Max. :98.0551 Max. :1.0534
## HV vhAscDesc_2015 vhAsc_2015 vhDesc_2015
## Min. :0.002783 Min. :0.01934 Min. :0.01458 Min. :0.01288
## 1st Qu.:0.030234 1st Qu.:0.02974 1st Qu.:0.02650 1st Qu.:0.02473
## Median :0.042002 Median :0.03321 Median :0.03261 Median :0.03209
## Mean :0.043745 Mean :0.03407 Mean :0.03742 Mean :0.03612
## 3rd Qu.:0.055124 3rd Qu.:0.03727 3rd Qu.:0.04172 3rd Qu.:0.04282
## Max. :0.118287 Max. :0.06863 Max. :0.20586 Max. :0.15697
## vvAscDesc_2015 vvAsc_2015 vvDesc_2015 SR_B2
## Min. :0.08023 Min. :0.04893 Min. :0.04136 Min. : 8037
## 1st Qu.:0.11088 1st Qu.:0.09140 1st Qu.:0.08664 1st Qu.: 8801
## Median :0.12287 Median :0.11742 Median :0.11746 Median : 9135
## Mean :0.13052 Mean :0.15415 Mean :0.15015 Mean : 9302
## 3rd Qu.:0.14157 3rd Qu.:0.16581 3rd Qu.:0.17863 3rd Qu.: 9473
## Max. :0.34069 Max. :0.93978 Max. :0.82425 Max. :15317
## brightness DVI EVI GNDVI
## Min. :22785 Min. :1853 Min. :-191.8228 Min. :0.08287
## 1st Qu.:25912 1st Qu.:2813 1st Qu.: 0.8833 1st Qu.:0.13523
## Median :27103 Median :3123 Median : 1.2357 Median :0.14680
## Mean :27411 Mean :3164 Mean : 0.8794 Mean :0.14790
## 3rd Qu.:28511 3rd Qu.:3455 3rd Qu.: 1.6432 3rd Qu.:0.15873
## Max. :39037 Max. :5209 Max. : 11.1121 Max. :0.22350
## SR_B3 greenness GVI LAI
## Min. : 8699 Min. :-4737.9 Min. :0.6350 Min. :-610.115
## 1st Qu.: 9666 1st Qu.:-2141.6 1st Qu.:0.7265 1st Qu.: 2.691
## Median :10056 Median :-1821.8 Median :0.7443 Median : 3.812
## Mean :10245 Mean :-1845.2 Mean :0.7434 Mean : 2.679
## 3rd Qu.:10565 3rd Qu.:-1479.9 3rd Qu.:0.7625 3rd Qu.: 5.107
## Max. :16042 Max. : 122.5 Max. :0.8524 Max. : 35.218
## NBR NDBI NDVI SR_B5
## Min. :0.02610 Min. :-0.19843 Min. :0.07217 Min. :11397
## 1st Qu.:0.07436 1st Qu.:-0.10663 1st Qu.:0.11658 1st Qu.:13050
## Median :0.08884 Median :-0.08884 Median :0.13068 Median :13646
## Mean :0.09088 Mean :-0.09088 Mean :0.13253 Mean :13760
## 3rd Qu.:0.10663 3rd Qu.:-0.07436 3rd Qu.:0.14603 3rd Qu.:14290
## Max. :0.19843 Max. :-0.02610 Max. :0.22641 Max. :18904
## NLI NRVI OSAVI RDVI
## Min. :0.0722 Min. :0.4120 Min. :0.1083 Min. :11.46
## 1st Qu.:0.1166 1st Qu.:0.6576 1st Qu.:0.1749 1st Qu.:18.14
## Median :0.1307 Median :0.6983 Median :0.1960 Median :20.24
## Mean :0.1325 Mean :0.6915 Mean :0.1988 Mean :20.45
## 3rd Qu.:0.1460 3rd Qu.:0.7351 3rd Qu.:0.2190 3rd Qu.:22.29
## Max. :0.2264 Max. :0.8372 Max. :0.3396 Max. :34.34
## SR_B4 RVI SAVI SR_B6
## Min. : 8621 Min. :0.6315 Min. :0.1083 Min. :10492
## 1st Qu.: 9918 1st Qu.:0.7453 1st Qu.:0.1749 1st Qu.:12621
## Median :10417 Median :0.7695 Median :0.1960 Median :13467
## Mean :10596 Mean :0.7673 Mean :0.1988 Mean :13519
## 3rd Qu.:11077 3rd Qu.:0.7917 3rd Qu.:0.2190 3rd Qu.:14400
## Max. :16451 Max. :0.8700 Max. :0.3396 Max. :17833
## SR_B7 SLAVI TVI wetness
## Min. : 9095 Min. :0.5361 Min. :0.7556 Min. :-5648
## 1st Qu.:10744 1st Qu.:0.5583 1st Qu.:0.7852 1st Qu.:-3697
## Median :11422 Median :0.5653 Median :0.7940 Median :-3209
## Mean :11496 Mean :0.5663 Mean :0.7951 Mean :-3269
## 3rd Qu.:12181 3rd Qu.:0.5730 3rd Qu.:0.8037 3rd Qu.:-2762
## Max. :14992 Max. :0.6132 Max. :0.8522 Max. :-1126
## DEM SLOPE ASPECT
## Min. : 502.5 Min. :0.04791 Min. :0.3384
## 1st Qu.: 682.3 1st Qu.:0.24013 1st Qu.:1.6807
## Median : 888.4 Median :0.30851 Median :2.8795
## Mean : 934.3 Mean :0.31211 Mean :2.9892
## 3rd Qu.:1209.3 3rd Qu.:0.37499 3rd Qu.:4.2867
## Max. :1860.2 Max. :0.67233 Max. :6.0826summary(test_PH)## FID OBJECTI N_Cintf Dnm_cm
## Min. : 19 Min. : 3024 Length:124 Min. :10.59
## 1st Qu.:1095 1st Qu.: 4102 Class :character 1st Qu.:13.48
## Median :1662 Median : 4669 Mode :character Median :15.39
## Mean :1466 Mean : 7100 Mean :17.22
## 3rd Qu.:1834 3rd Qu.: 4840 3rd Qu.:20.02
## Max. :2830 Max. :46678 Max. :37.25
## Npies_h AB_m2_h Dmc_cm Ho_m
## Min. : 20.05 Min. : 0.1766 Min. :10.59 Min. : 3.540
## 1st Qu.: 240.60 1st Qu.: 4.3174 1st Qu.:13.56 1st Qu.: 6.803
## Median : 380.95 Median : 7.2154 Median :15.70 Median : 7.832
## Mean : 429.48 Mean :10.5365 Mean :17.46 Mean : 8.915
## 3rd Qu.: 541.34 3rd Qu.:14.4403 3rd Qu.:20.52 3rd Qu.:10.945
## Max. :1303.23 Max. :42.1995 Max. :37.25 Max. :15.788
## Bttl_Kg MgC.pr_ Ctt_MC. HH
## Min. : 131.6 Min. :0.0126 Min. : 0.2527 Min. :0.0279
## 1st Qu.: 221.3 1st Qu.:0.3353 1st Qu.: 6.7231 1st Qu.:0.1146
## Median : 325.4 Median :0.6031 Median :12.0924 Median :0.1608
## Mean : 487.2 Mean :1.0220 Mean :20.4917 Mean :0.1678
## 3rd Qu.: 647.2 3rd Qu.:1.3774 3rd Qu.:27.6164 3rd Qu.:0.2005
## Max. :2294.8 Max. :4.4170 Max. :88.5609 Max. :0.6073
## HV vhAscDesc_2015 vhAsc_2015 vhDesc_2015
## Min. :0.005694 Min. :0.02196 Min. :0.01304 Min. :0.01436
## 1st Qu.:0.029249 1st Qu.:0.03081 1st Qu.:0.02640 1st Qu.:0.02480
## Median :0.044233 Median :0.03326 Median :0.03237 Median :0.03324
## Mean :0.045533 Mean :0.03461 Mean :0.03885 Mean :0.03675
## 3rd Qu.:0.060632 3rd Qu.:0.03714 3rd Qu.:0.04293 3rd Qu.:0.04212
## Max. :0.118746 Max. :0.05848 Max. :0.11844 Max. :0.14526
## vvAscDesc_2015 vvAsc_2015 vvDesc_2015 SR_B2
## Min. :0.08491 Min. :0.04756 Min. :0.04328 Min. : 8031
## 1st Qu.:0.10966 1st Qu.:0.08980 1st Qu.:0.08487 1st Qu.: 8677
## Median :0.12232 Median :0.12272 Median :0.12038 Median : 9017
## Mean :0.13174 Mean :0.16257 Mean :0.15160 Mean : 9217
## 3rd Qu.:0.13799 3rd Qu.:0.16565 3rd Qu.:0.17280 3rd Qu.: 9505
## Max. :0.34326 Max. :0.97181 Max. :0.78906 Max. :14259
## brightness DVI EVI GNDVI
## Min. :22467 Min. :2090 Min. :-39.8404 Min. :0.1064
## 1st Qu.:25578 1st Qu.:2872 1st Qu.: 0.8391 1st Qu.:0.1349
## Median :26614 Median :3105 Median : 1.3940 Median :0.1468
## Mean :27190 Mean :3182 Mean : 0.9492 Mean :0.1483
## 3rd Qu.:28611 3rd Qu.:3472 3rd Qu.: 1.7009 3rd Qu.:0.1601
## Max. :36922 Max. :4838 Max. : 4.5486 Max. :0.2060
## SR_B3 greenness GVI LAI
## Min. : 8717 Min. :-4320.8 Min. :0.6584 Min. :-126.810
## 1st Qu.: 9548 1st Qu.:-2201.5 1st Qu.:0.7241 1st Qu.: 2.550
## Median : 9918 Median :-1788.2 Median :0.7442 Median : 4.315
## Mean :10158 Mean :-1799.3 Mean :0.7427 Mean : 2.900
## 3rd Qu.:10679 3rd Qu.:-1458.7 3rd Qu.:0.7626 3rd Qu.: 5.291
## Max. :14570 Max. : -408.6 Max. :0.8106 Max. : 14.347
## NBR NDBI NDVI SR_B5
## Min. :0.02585 Min. :-0.15145 Min. :0.07982 Min. :11292
## 1st Qu.:0.07343 1st Qu.:-0.10852 1st Qu.:0.11677 1st Qu.:12869
## Median :0.09226 Median :-0.09226 Median :0.13194 Median :13479
## Mean :0.09179 Mean :-0.09179 Mean :0.13416 Mean :13664
## 3rd Qu.:0.10852 3rd Qu.:-0.07343 3rd Qu.:0.14789 3rd Qu.:14215
## Max. :0.15145 Max. :-0.02585 Max. :0.19602 Max. :17850
## NLI NRVI OSAVI RDVI
## Min. :0.07984 Min. :0.5114 Min. :0.1197 Min. :13.01
## 1st Qu.:0.11677 1st Qu.:0.6523 1st Qu.:0.1752 1st Qu.:18.44
## Median :0.13195 Median :0.6954 Median :0.1979 Median :20.03
## Mean :0.13416 Mean :0.6870 Mean :0.2012 Mean :20.63
## 3rd Qu.:0.14790 3rd Qu.:0.7341 3rd Qu.:0.2218 3rd Qu.:22.50
## Max. :0.19600 Max. :0.8257 Max. :0.2940 Max. :30.44
## SR_B4 RVI SAVI SR_B6
## Min. : 8625 Min. :0.6725 Min. :0.1197 Min. :10501
## 1st Qu.: 9760 1st Qu.:0.7426 1st Qu.:0.1752 1st Qu.:12489
## Median :10257 Median :0.7671 Median :0.1979 Median :13123
## Mean :10482 Mean :0.7648 Mean :0.2012 Mean :13428
## 3rd Qu.:11163 3rd Qu.:0.7919 3rd Qu.:0.2218 3rd Qu.:14372
## Max. :14774 Max. :0.8527 Max. :0.2940 Max. :18685
## SR_B7 SLAVI TVI wetness
## Min. : 9161 Min. :0.5399 Min. :0.7614 Min. :-5812
## 1st Qu.:10638 1st Qu.:0.5584 1st Qu.:0.7852 1st Qu.:-3659
## Median :11186 Median :0.5660 Median :0.7949 Median :-3149
## Mean :11405 Mean :0.5671 Mean :0.7961 Mean :-3263
## 3rd Qu.:12131 3rd Qu.:0.5739 3rd Qu.:0.8049 3rd Qu.:-2769
## Max. :15743 Max. :0.5980 Max. :0.8342 Max. :-1398
## DEM SLOPE ASPECT
## Min. : 527.1 Min. :0.1030 Min. :0.551
## 1st Qu.: 685.4 1st Qu.:0.2574 1st Qu.:1.731
## Median : 809.7 Median :0.3325 Median :2.866
## Mean : 916.5 Mean :0.3316 Mean :3.062
## 3rd Qu.:1164.1 3rd Qu.:0.3977 3rd Qu.:4.415
## Max. :1397.4 Max. :0.5722 Max. :5.906#
# Dividir los datos en un conjunto de entrenamiento (70%) y un conjunto de prueba (30%)
set.seed(123)
train_index <- createDataPartition(y =DS_PP$Ctt_MC., p = 0.8, list = FALSE)
train_PP <-DS_PP[train_index, ]
test_PP <-DS_PP[-train_index, ]
summary(train_PP)## FID OBJECTI N_Cintf Dnm_cm
## Min. : 8.0 Min. : 3007 Length:575 Min. :12.32
## 1st Qu.: 689.5 1st Qu.: 3696 Class :character 1st Qu.:19.74
## Median :1245.0 Median : 4252 Mode :character Median :23.46
## Mean :1231.3 Mean : 7339 Mean :23.40
## 3rd Qu.:1556.5 3rd Qu.: 4564 3rd Qu.:26.68
## Max. :2829.0 Max. :46677 Max. :58.22
## Npies_h AB_m2_h Dmc_cm Ho_m
## Min. : 14.15 Min. : 0.3187 Min. :12.42 Min. : 5.457
## 1st Qu.: 252.34 1st Qu.: 9.6089 1st Qu.:20.03 1st Qu.: 9.792
## Median : 501.24 Median :21.2109 Median :23.79 Median :11.429
## Mean : 532.77 Mean :23.6669 Mean :23.74 Mean :11.296
## 3rd Qu.: 761.89 3rd Qu.:35.8138 3rd Qu.:27.05 3rd Qu.:12.794
## Max. :1784.43 Max. :66.9574 Max. :58.23 Max. :21.330
## Bttl_Kg MgC.pr_ Ctt_MC. HH
## Min. : 27.29 Min. :0.01862 Min. : 0.3734 Min. :0.02432
## 1st Qu.: 353.15 1st Qu.:0.77723 1st Qu.: 15.5836 1st Qu.:0.13050
## Median : 563.79 Median :1.89458 Median : 37.9866 Median :0.17496
## Mean : 684.88 Mean :2.13374 Mean : 42.7818 Mean :0.18461
## 3rd Qu.: 937.98 3rd Qu.:3.27723 3rd Qu.: 65.7088 3rd Qu.:0.22455
## Max. :3744.09 Max. :5.71134 Max. :114.5130 Max. :1.03525
## HV vhAscDesc_2015 vhAsc_2015 vhDesc_2015
## Min. :0.004446 Min. :0.01754 Min. :0.009292 Min. :0.01017
## 1st Qu.:0.033228 1st Qu.:0.02615 1st Qu.:0.021690 1st Qu.:0.01997
## Median :0.050045 Median :0.02902 Median :0.027518 Median :0.02705
## Mean :0.051441 Mean :0.02934 Mean :0.033008 Mean :0.03107
## 3rd Qu.:0.067803 3rd Qu.:0.03211 3rd Qu.:0.040275 3rd Qu.:0.03888
## Max. :0.152579 Max. :0.06309 Max. :0.125276 Max. :0.10816
## vvAscDesc_2015 vvAsc_2015 vvDesc_2015 SR_B2
## Min. :0.07440 Min. :0.03534 Min. :0.03114 Min. : 7925
## 1st Qu.:0.09466 1st Qu.:0.07204 1st Qu.:0.06728 1st Qu.: 8393
## Median :0.10690 Median :0.09711 Median :0.10055 Median : 8790
## Mean :0.11324 Mean :0.14376 Mean :0.13354 Mean : 9161
## 3rd Qu.:0.12368 3rd Qu.:0.16553 3rd Qu.:0.15615 3rd Qu.: 9311
## Max. :0.31366 Max. :2.86281 Max. :0.99251 Max. :17183
## brightness DVI EVI GNDVI
## Min. :22467 Min. :1799 Min. :-75.0934 Min. :0.1054
## 1st Qu.:25457 1st Qu.:3517 1st Qu.: 0.9531 1st Qu.:0.1619
## Median :26799 Median :4345 Median : 1.3764 Median :0.1939
## Mean :27646 Mean :4334 Mean : 1.1835 Mean :0.1917
## 3rd Qu.:28676 3rd Qu.:5133 3rd Qu.: 1.9397 3rd Qu.:0.2223
## Max. :45167 Max. :6422 Max. : 11.6576 Max. :0.2603
## SR_B3 greenness GVI LAI
## Min. : 8530 Min. :-5204.90 Min. :0.5870 Min. :-238.915
## 1st Qu.: 9194 1st Qu.:-1698.64 1st Qu.:0.6366 1st Qu.: 2.913
## Median : 9723 Median : -833.85 Median :0.6759 Median : 4.259
## Mean :10110 Mean : -925.58 Mean :0.6807 Mean : 3.645
## 3rd Qu.:10469 3rd Qu.: 17.42 3rd Qu.:0.7223 3rd Qu.: 6.050
## Max. :17903 Max. : 1000.74 Max. :0.8097 Max. : 36.953
## NBR NDBI NDVI SR_B5
## Min. :0.04092 Min. :-0.24051 Min. :0.07198 Min. :11292
## 1st Qu.:0.10258 1st Qu.:-0.17892 1st Qu.:0.14045 1st Qu.:14000
## Median :0.13733 Median :-0.13733 Median :0.17521 Median :14515
## Mean :0.14016 Mean :-0.14016 Mean :0.17679 Mean :14788
## 3rd Qu.:0.17892 3rd Qu.:-0.10258 3rd Qu.:0.21617 3rd Qu.:15267
## Max. :0.24051 Max. :-0.04092 Max. :0.26365 Max. :21711
## NLI NRVI OSAVI RDVI
## Min. :0.07195 Min. :0.2835 Min. :0.1080 Min. :11.38
## 1st Qu.:0.14045 1st Qu.:0.4438 1st Qu.:0.2107 1st Qu.:22.19
## Median :0.17521 Median :0.5722 Median :0.2628 Median :27.71
## Mean :0.17679 Mean :0.5607 Mean :0.2652 Mean :27.64
## 3rd Qu.:0.21618 3rd Qu.:0.6699 3rd Qu.:0.3243 3rd Qu.:33.37
## Max. :0.26366 Max. :0.8446 Max. :0.3955 Max. :41.14
## SR_B4 RVI SAVI SR_B6
## Min. : 8408 Min. :0.5828 Min. :0.1080 Min. : 9947
## 1st Qu.: 9318 1st Qu.:0.6458 1st Qu.:0.2107 1st Qu.:11818
## Median :10112 Median :0.7031 Median :0.2628 Median :12900
## Mean :10454 Mean :0.7031 Mean :0.2652 Mean :13232
## 3rd Qu.:11081 3rd Qu.:0.7559 3rd Qu.:0.3243 3rd Qu.:14260
## Max. :18647 Max. :0.8659 Max. :0.3955 Max. :21085
## SR_B7 SLAVI TVI wetness
## Min. : 8724 Min. :0.5360 Min. :0.7563 Min. :-6257.8
## 1st Qu.:10040 1st Qu.:0.5702 1st Qu.:0.8000 1st Qu.:-3380.6
## Median :10910 Median :0.5876 Median :0.8215 Median :-2690.7
## Mean :11236 Mean :0.5884 Mean :0.8220 Mean :-2689.6
## 3rd Qu.:12055 3rd Qu.:0.6081 3rd Qu.:0.8459 3rd Qu.:-1895.6
## Max. :18274 Max. :0.6318 Max. :0.8739 Max. : -547.3
## DEM SLOPE ASPECT
## Min. : 903.2 Min. :0.1086 Min. :0.1948
## 1st Qu.:1339.2 1st Qu.:0.2496 1st Qu.:1.5709
## Median :1481.1 Median :0.3087 Median :2.7044
## Mean :1453.4 Mean :0.3090 Mean :2.9402
## 3rd Qu.:1572.4 3rd Qu.:0.3648 3rd Qu.:4.3246
## Max. :1918.2 Max. :0.6088 Max. :6.0094summary(test_PP)## FID OBJECTI N_Cintf Dnm_cm
## Min. : 10.0 Min. : 3015 Length:140 Min. :11.12
## 1st Qu.: 851.5 1st Qu.: 3858 Class :character 1st Qu.:19.70
## Median :1290.5 Median : 4298 Mode :character Median :23.61
## Mean :1304.3 Mean : 7500 Mean :23.21
## 3rd Qu.:1581.8 3rd Qu.: 4589 3rd Qu.:26.76
## Max. :2785.0 Max. :46317 Max. :37.09
## Npies_h AB_m2_h Dmc_cm Ho_m
## Min. : 20.05 Min. : 0.3896 Min. :11.12 Min. : 4.577
## 1st Qu.: 260.65 1st Qu.:10.8463 1st Qu.:19.89 1st Qu.: 9.825
## Median : 481.19 Median :20.7148 Median :23.87 Median :11.392
## Mean : 532.40 Mean :23.4803 Mean :23.52 Mean :11.160
## 3rd Qu.: 781.94 3rd Qu.:36.2221 3rd Qu.:27.11 3rd Qu.:12.707
## Max. :1846.20 Max. :63.6446 Max. :37.09 Max. :15.777
## Bttl_Kg MgC.pr_ Ctt_MC. HH
## Min. : 4.819 Min. :0.01961 Min. : 0.3931 Min. :0.02729
## 1st Qu.: 352.909 1st Qu.:0.78962 1st Qu.: 15.8320 1st Qu.:0.12521
## Median : 573.988 Median :1.89119 Median : 37.9186 Median :0.17741
## Mean : 640.514 Mean :2.09958 Mean : 42.0968 Mean :0.18351
## 3rd Qu.: 909.323 3rd Qu.:3.24580 3rd Qu.: 65.0787 3rd Qu.:0.22511
## Max. :2165.582 Max. :5.90087 Max. :118.3131 Max. :1.00861
## HV vhAscDesc_2015 vhAsc_2015 vhDesc_2015
## Min. :0.005401 Min. :0.01979 Min. :0.01043 Min. :0.01234
## 1st Qu.:0.034576 1st Qu.:0.02582 1st Qu.:0.02143 1st Qu.:0.02068
## Median :0.053534 Median :0.02832 Median :0.02748 Median :0.02782
## Mean :0.052972 Mean :0.02907 Mean :0.03162 Mean :0.03047
## 3rd Qu.:0.068736 3rd Qu.:0.03179 3rd Qu.:0.03935 3rd Qu.:0.03799
## Max. :0.122027 Max. :0.06019 Max. :0.10879 Max. :0.07413
## vvAscDesc_2015 vvAsc_2015 vvDesc_2015 SR_B2
## Min. :0.07594 Min. :0.04594 Min. :0.03886 Min. : 7916
## 1st Qu.:0.09505 1st Qu.:0.07325 1st Qu.:0.06855 1st Qu.: 8397
## Median :0.10770 Median :0.09796 Median :0.10066 Median : 8742
## Mean :0.11109 Mean :0.12863 Mean :0.12838 Mean : 8886
## 3rd Qu.:0.12261 3rd Qu.:0.15800 3rd Qu.:0.16031 3rd Qu.: 9216
## Max. :0.24180 Max. :0.57784 Max. :0.46730 Max. :13953
## brightness DVI EVI GNDVI
## Min. :22785 Min. :1751 Min. :-19.2800 Min. :0.1026
## 1st Qu.:25662 1st Qu.:3633 1st Qu.: 0.9579 1st Qu.:0.1680
## Median :26851 Median :4567 Median : 1.3776 Median :0.2036
## Mean :27118 Mean :4436 Mean : 1.3499 Mean :0.1958
## 3rd Qu.:28391 3rd Qu.:5417 3rd Qu.: 1.8295 3rd Qu.:0.2294
## Max. :39184 Max. :6090 Max. : 10.9927 Max. :0.2515
## SR_B3 greenness GVI LAI
## Min. : 8535 Min. :-3202.7 Min. :0.5982 Min. :-61.428
## 1st Qu.: 9239 1st Qu.:-1440.4 1st Qu.:0.6282 1st Qu.: 2.928
## Median : 9649 Median : -560.8 Median :0.6619 Median : 4.263
## Mean : 9854 Mean : -739.4 Mean :0.6747 Mean : 4.175
## 3rd Qu.:10268 3rd Qu.: 121.5 3rd Qu.:0.7126 3rd Qu.: 5.700
## Max. :14913 Max. : 706.1 Max. :0.8141 Max. : 34.839
## NBR NDBI NDVI SR_B5
## Min. :0.03846 Min. :-0.23880 Min. :0.06742 Min. :11397
## 1st Qu.:0.10567 1st Qu.:-0.18236 1st Qu.:0.14796 1st Qu.:14054
## Median :0.14478 Median :-0.14478 Median :0.18604 Median :14581
## Mean :0.14262 Mean :-0.14262 Mean :0.18098 Mean :14621
## 3rd Qu.:0.18236 3rd Qu.:-0.10567 3rd Qu.:0.22026 3rd Qu.:15037
## Max. :0.23880 Max. :-0.03846 Max. :0.25315 Max. :19854
## NLI NRVI OSAVI RDVI
## Min. :0.06742 Min. :0.3217 Min. :0.1011 Min. :10.87
## 1st Qu.:0.14798 1st Qu.:0.4319 1st Qu.:0.2219 1st Qu.:22.97
## Median :0.18605 Median :0.5378 Median :0.2791 Median :29.21
## Mean :0.18098 Mean :0.5485 Mean :0.2715 Mean :28.31
## 3rd Qu.:0.22026 3rd Qu.:0.6495 3rd Qu.:0.3304 3rd Qu.:34.42
## Max. :0.25315 Max. :0.8551 Max. :0.3797 Max. :38.88
## SR_B4 RVI SAVI SR_B6
## Min. : 8476 Min. :0.5961 Min. :0.1011 Min. :10006
## 1st Qu.: 9345 1st Qu.:0.6395 1st Qu.:0.2219 1st Qu.:11885
## Median :10022 Median :0.6866 Median :0.2791 Median :12863
## Mean :10185 Mean :0.6968 Mean :0.2715 Mean :13010
## 3rd Qu.:10830 3rd Qu.:0.7434 3rd Qu.:0.3304 3rd Qu.:13970
## Max. :15557 Max. :0.8739 Max. :0.3797 Max. :18500
## SR_B7 SLAVI TVI wetness
## Min. : 8732 Min. :0.5337 Min. :0.7532 Min. :-5906.1
## 1st Qu.:10037 1st Qu.:0.5740 1st Qu.:0.8048 1st Qu.:-3306.4
## Median :10970 Median :0.5930 Median :0.8282 Median :-2695.4
## Mean :11024 Mean :0.5905 Mean :0.8246 Mean :-2673.1
## 3rd Qu.:11807 3rd Qu.:0.6101 3rd Qu.:0.8486 3rd Qu.:-1896.1
## Max. :15629 Max. :0.6266 Max. :0.8678 Max. : -721.9
## DEM SLOPE ASPECT
## Min. : 909.3 Min. :0.1079 Min. :0.5926
## 1st Qu.:1387.6 1st Qu.:0.2301 1st Qu.:1.7874
## Median :1504.6 Median :0.3154 Median :2.8702
## Mean :1458.8 Mean :0.3003 Mean :2.9709
## 3rd Qu.:1566.4 3rd Qu.:0.3610 3rd Qu.:4.1765
## Max. :1722.8 Max. :0.4739 Max. :6.0102fitControl <- trainControl(method = "repeatedcv", number=10, repeats=5,
#classProbs = TRUE,
returnResamp="all",
returnData = TRUE,
savePredictions=TRUE)MODELO Pinus Pinaster BIOMASA
model_PP_AGB <- train (Ctt_MC. ~
HH + HV + vhDesc_2015 + vvAscDesc_2015+ vvAsc_2015
+ vvDesc_2015 + SR_B2 + DVI + LAI + NDBI
+ DEM + SLOPE + ASPECT
,data=train_PP, method="rf",
trControl=fitControl,
prox=TRUE,
fitBest = TRUE,
)
model_PP_AGB## Random Forest
##
## 575 samples
## 13 predictor
##
## No pre-processing
## Resampling: Cross-Validated (10 fold, repeated 5 times)
## Summary of sample sizes: 519, 517, 517, 518, 518, 516, ...
## Resampling results across tuning parameters:
##
## mtry RMSE Rsquared MAE
## 2 19.63748 0.6081623 15.25107
## 7 19.55525 0.6090764 14.94387
## 13 19.66404 0.6048545 14.94305
##
## RMSE was used to select the optimal model using the smallest value.
## The final value used for the model was mtry = 7.hacemos una evaluación con el set de testeo
Predict_model_PP_AGB <- predict(model_PP_AGB, newdata=test_PP)
Observed_ctt <- test_PP$Ctt_MC.
tmp_ctt <- data.frame(Predict_model_PP_AGB, Observed_ctt)
ev<- caret::postResample(test_PP$Ctt_MC., Predict_model_PP_AGB)
ev## RMSE Rsquared MAE
## 18.6565594 0.6325228 13.4633422por ultimo graficamos el modelo observado vs el modelo predicho
m1_PN <- tmp_ctt %>% dplyr::select( Predict_model_PP_AGB, Observed_ctt) %>%
pivot_longer(cols = -Observed_ctt) %>%
ggplot(aes(x = Observed_ctt, y = value)) + geom_point()+
stat_smooth(aes(), method="lm", formula=y ~ x) +theme_bw()+
ylab("Model Pinud pinaster (Mg·ha−1)")+ xlab("observed (Mg·ha−1)")+
ggtitle("MODEL RF PP VS OBSERVED",subtitle = "R2=0.49 - RSME= 21.43 - %RSME= 16.90%")+
geom_abline(intercept=0, slope=1, lwd=1, linetype=2, color="red")
m1_PN 
MgC_PP<- predict(rasters1, model_PP_AGB)
MgC_PP_mask <- mask(MgC_PP, pinus_PP)
plot(MgC_PP_mask)
writeRaster(MgC_PP_mask,filename = paste ("D:/cap_13_Radar/RESULTADOS/","MgC_PP_2015"), bylayer=TRUE, format="GTiff", overwrite=TRUE, sep="")MODELO Pinus Halepensis BIOMASA
model_PH_AGB <- train (Ctt_MC. ~
HH + HV + vhDesc_2015 + vvAscDesc_2015+ vvAsc_2015
+ vvDesc_2015 + SR_B2 + DVI + LAI + NDBI
+ DEM + SLOPE + ASPECT
,data=train_PH, method="rf",
trControl=fitControl,
prox=TRUE,
fitBest = TRUE,
)
model_PH_AGB## Random Forest
##
## 506 samples
## 13 predictor
##
## No pre-processing
## Resampling: Cross-Validated (10 fold, repeated 5 times)
## Summary of sample sizes: 454, 455, 456, 455, 455, 457, ...
## Resampling results across tuning parameters:
##
## mtry RMSE Rsquared MAE
## 2 13.32181 0.5018863 9.433272
## 7 13.60205 0.4725944 9.475126
## 13 13.79335 0.4581861 9.552726
##
## RMSE was used to select the optimal model using the smallest value.
## The final value used for the model was mtry = 2.Predict_model_PH_AGB <- predict(model_PH_AGB, newdata=test_PH)
Observed_ctt <- test_PH$Ctt_MC.
tmp_ctt <- data.frame(Predict_model_PH_AGB, Observed_ctt)
ev<- caret::postResample(test_PH$Ctt_MC. ,Predict_model_PH_AGB)
ev## RMSE Rsquared MAE
## 13.1817939 0.5657525 9.3199807m1_PN <- tmp_ctt %>% dplyr::select( Predict_model_PH_AGB, Observed_ctt) %>%
pivot_longer(cols = -Observed_ctt) %>%
ggplot(aes(x = Observed_ctt, y = value)) + geom_point()+
stat_smooth(aes(), method="lm", formula=y ~ x) +theme_bw()+
ylab("Model RF Wt (Mg·ha−1)")+ xlab("observed (Mg·ha−1)")+
ggtitle("MODEL RF Wt VS OBSERVED",subtitle = "R2=0.49 - RSME= 21.43 - %RSME= 16.90%")+
geom_abline(intercept=0, slope=1, lwd=1, linetype=2, color="red")
m1_PN 
MgC_PH<- predict(rasters1, model_PH_AGB)
MgC_PH_mask <- mask(MgC_PH, pinus_PH)
plot(MgC_PH_mask)
writeRaster(MgC_PH_mask,filename = paste ("D:/cap_13_Radar/RESULTADOS/","MgC_PH_2015"), bylayer=TRUE, format="GTiff", overwrite=TRUE, sep="")